Thread by @GrahamNeary

I've gone deep into the mortality numbers for Ireland and Covid-19. This is what I've learned.

Thread.

The death toll in Ireland from Covid-19 is officially 1,773. But this includes people who died from other things and people who were presumed to have it.

We need to do better, to find out how many people were really killed by Covid-19.

Analysts for the Irish government counted death notices online and reckoned there were between 851 and 1290 excess deaths by mid-June, with the middle estimate being 1,072.

Instead of using online death notices, I have sought out and gained access to the latest official death registration data, as of the end of July 2020.

Here it is.

CC @FatEmperor @Niall_Boylan @DavQuinn @HealthFreedomIE

Here are all of the monthly results from January to May for each year, ordered from largest to smallest. April 2020 and May 2020 are in red:

Here are the cumulative deaths from January to May for each year:

Two big oversights are:

1) Not all deaths for 2020 are registered yet.

Officially you have three months to register a death, but I can see from previous snapshots that deaths continue to be registered after three months.

2) The population is growing.

My solutions:

1) I use differences between the end-June database and the end-July database to estimate likely future growth of 2020 numbers for Jan to May, through the end of October.

2) I divide all numbers by the population for each year, to get the monthly mortality rates.

Results:

I estimate that April 2020 was about as dangerous as Jan 2018 and Jan 2017 (flu seasons).

And May 2020 was about as dangerous as any other April or May.

So far, 2020 looks like it has been less dangerous than 2016, 2017 and 2018, based on deaths until May.

It is only looking more dangerous than 2017 and 2019.

Not exactly the message you'd pick up from the mainstream media, is it?

What makes April 2020 different? It's certainly shaping up to have been more dangerous than other Aprils:

And it was: the health system was shut down for many treatments, inpatients were asked to leave early, and many were too scared to attend A&E or their GP.

For England, the ONS estimated 16,000 non-Covid excess deaths were caused by lockdown.

.0062 deaths per 1000 people per day

The best I can do for an initial estimate of deaths in Ireland caused by lockdown is to translate this ONS figure across.

Resulting estimate: 916 deaths caused by lockdown in Ireland in April 2020.

This wouldn't leave very many excess deaths available for Covid and other causes

I acknowledge there are large gaps in my analysis. Detailed research (and the passing of time) will provide in-depth explanations for many of the above issues.

However, I don't see why my estimates would not be in the "ballpark" of truth. Please share with me your corrections.

My conclusions:

1) data does not yet support the claim that an extremely deadly virus (e.g. more dangerous than flu) was present and active in Ireland from March 2020.

2) deaths caused immediately by lockdown might have been significantly higher than deaths caused by the virus.

Here's a YouTube video talking through all of the above:

Should be: less dangerous than 2015, 2016, 2018.